Re: [lwip-users] STM32F7 Project

[Noam Weissman]

Hi,

I am using FreeRTOS + LwIP 1.41 on STM32F4xx and STM32F2xx for a few years now.

We do not use the cube as we have lots of base code that uses the older ST library.

The benefits of using the cube is that it should be easy to port to another OS or TCP stack.

The penalty is performance. As they use lots of wrappers to do this. I do not think that it

is so critical but this is how its being done.

We are planing to move to the 765/767 in the near time.

Do remember that you do not have really 512K RAM, 128K is internal like the CCM in 427

It is not so simple to use it. Not to complicated but not trivial. I use the CCM for heap in

case we use SSL.

Good luck,

Noam.

---

From: lwip-users <lwip-users-bounces+address@hidden> on behalf of Kenny Koller <address@hidden>
Sent: Sunday, October 2, 2016 11:52 PM
To: LWIP
Subject: [lwip-users] STM32F7 Project

 

Hi Everyone,

I have a code base running FreeRTOS on an STM32F767 (512 kB RAM) and I am preparing to integrate LWIP. Currently I am using the Nucleo-F767 board and our PCB will be using the same PHY and RMII.

I do not use the ST Cube HAL but plan to use the Cube generated code as a reference. There is also an ST application note for the STM32F4 and I spent some time reading the EMAC portion the reference manual.

My thinking is that I will begin by getting the Cube LWIP up and running and then port to 2.0 following the LWIP guide. I'm using the latest Segger Embedded Studio on a Mac (ARM GCC 5). Does anyone see a problem with this approach? Caveats or suggestions?

Also, I noticed that ST CubeMX (4.16.1) generated code is LWIP 1.5 but I didn't see a tag in the git repository for this version:

Silverback: ack LWIP_VERSION

src/include/lwip/init.h

42:#define LWIP_VERSION_MAJOR      1U

44:#define LWIP_VERSION_MINOR      5U

46:#define LWIP_VERSION_REVISION   0U

50:#define LWIP_VERSION_RC         0U

I'm looking forward to using the stack.

Thanks,

Kenny